gsd-thread

To persist context threads across sessions, you can create lightweight thread stores under a planning directory. This approach saves notes, goals, and next steps as plain text files, allowing you to resume work weeks later with full context.

Managing long-running research context is done by creating persistent threads that hold analysis data. You list, resume, or close these threads, ensuring cross-session knowledge is preserved without manual copying.

You create and resume persistent threads by invoking a skill with a task description, which sanitizes slug names and updates frontmatter. Resume mode reloads the specific thread content for immediate continuation.

Thread persistence handles plain text notes by preserving them as-is, while structured data requires DATA_START and DATA_END boundaries. The system updates frontmatter and commits changes with descriptive messages.

Yes, you can track cross-session planning threads using status and list modes. These modes display thread summaries and current states, helping you identify which threads are active and which are closed.

Using plain text files for session context persistence limits structured data parsing unless you explicitly define DATA_START and DATA_END boundaries. Thread content remains unstructured text, requiring manual formatting for complex data.